Abstract: A system for mapping and localization includes first and second robotic devices and fiducial tags placed at predetermined positions of an environment. The quality of the first robotic device is better than the second robotic device. The first robotic device prepares a high resolution multilayer map, and the second robotic device uses the multilayer map for fast and accurate localization. The first robotic device has a visual sensor and a light detection and ranging (LIDAR) device, and is configured to prepare the multilayer map by: constructing a feature point cloud layer of the environment using images captured by the visual sensor; building a tag layer comprising identifications and locations of the tags; constructing an occupancy map using LIDAR scans; and forming the multilayer map having the feature point cloud layer, the tag layer, and the occupancy map layer, which share the same coordinate system.

Abstract: A system and method for constructing and updating a two dimensional (2D) grid cell map based on depth images and RGB images. The method includes: providing an RGB image and a depth image by a sensor of a computing device; determining scores of pixels in the RGB image representing possibility of the pixels being located in a ground plane; determining normal directions of pixels in the depth image; determining the ground plane based on the scores and the normal directions; projecting the pixels in the depth images to the ground plane; and generating and updating the 2D grid cell map by defining the ground plane into grid cells and attributing a color descriptor to each grid cell. The color descriptor of each grid cell is calculated based on the projected pixels within that grid cell.

Abstract: A method for performing localization and mapping of a mapping device. The method includes: capturing images of environment by a visual sensor of the mapping device, and processing the captured images to obtain a set of high-resolution images and a set of low-resolution images; determining motion-scale measurement of the mapping device using an inertial sensor and an odometer; generating, by a controller of the mapping device, a first mask corresponding to dynamic features in the low-resolution images; extracting features from the high-resolution images; recovering, by the controller, poses of the mapping device based on the high-resolution images and the motion-scale measurements; filtering out, by the controller, the features from the high-resolution images that correspond to the first mask, to obtain filtered features; and performing, by the controller, localization and mapping using the poses and the filtered features.

Abstract: A system and method for constructing and updating a two dimensional (2D) grid cell map based on depth images and RGB images. The method includes: providing an RGB image and a depth image by a sensor of a computing device; determining scores of pixels in the RGB image representing possibility of the pixels being located in a ground plane; determining normal directions of pixels in the depth image; determining the ground plane based on the scores and the normal directions; projecting the pixels in the depth images to the ground plane; and generating and updating the 2D grid cell map by defining the ground plane into grid cells and attributing a color descriptor to each grid cell. The color descriptor of each grid cell is calculated based on the projected pixels within that grid cell.

Abstract: A system for mapping and localization includes first and second robotic devices and fiducial tags placed at predetermined positions of an environment. The quality of the first robotic device is better than the second robotic device. The first robotic device prepares a high resolution multilayer map, and the second robotic device uses the multilayer map for fast and accurate localization. The first robotic device has a visual sensor and a light detection and ranging (LIDAR) device, and is configured to prepare the multilayer map by: constructing a feature point cloud layer of the environment using images captured by the visual sensor; building a tag layer comprising identifications and locations of the tags; constructing an occupancy map using LIDAR scans; and forming the multilayer map having the feature point cloud layer, the tag layer, and the occupancy map layer, which share the same coordinate system.

Abstract: A method for performing localization and mapping of a mapping device. The method includes: capturing images of environment by a visual sensor of the mapping device, and processing the captured images to obtain a set of high-resolution images and a set of low-resolution images; determining motion-scale measurement of the mapping device using an inertial sensor and an odometer; generating, by a controller of the mapping device, a first mask corresponding to dynamic features in the low-resolution images; extracting features from the high-resolution images; recovering, by the controller, poses of the mapping device based on the high-resolution images and the motion-scale measurements; filtering out, by the controller, the features from the high-resolution images that correspond to the first mask, to obtain filtered features; and performing, by the controller, localization and mapping using the poses and the filtered features.